Vincent Yan Fu Tan

Graduate Student, EECS Department, MIT.
Member of the Stochastic Systems Group (SSG), Laboratory for Information and Decision Systems (LIDS).

MIT, Room 32-D570
77 Mass. Ave
Cambridge, MA 02139

Phone: (617) 253-3816

Email: vtan at mit dot edu

 

Back in Cambridge for Huili’s MA graduation.

 

Brief Biography

Vincent Tan is a third-year graduate student at Laboratory for Information and Decision Systems (LIDS) in the Department of EECS at MIT. His research interests are in the broad areas of statistical signal processing, graphical models and convex optimization. He is affiliated to the Stochastic Systems Group (SSG) led by Prof. Alan Willsky.

Vincent was an undergraduate in Electrical and Information Sciences (EIST) at Sidney Sussex College in Cambridge University. He worked on his Masters project at the Signal Processing Laboratory in the Engineering Department (CUED) under Dr. Cédric Févotte and received the Charles Lamb Prize for being the top student in EIST in 2005. He spent his junior year at the Massachusetts Institute of Technology (MIT) on the Cambridge-MIT (CMI) Undergraduate Exchange Program.

·        In the summer of 2009, Vincent will be an intern at Microsoft Research.

·        In the summer of 2008, Vincent was an intern at the Machine Learning and Perception (MLP) group at Microsoft Research Cambridge (MSR-C).

·        From 2006 to 2007, Vincent worked at the Institute for Infocomm Research (I2R), a research institute under A*STAR.

·        From 2005 to 2006, Vincent worked as a research engineer at the Defence Science Organisation (DSO) National Laboratories in Singapore.

·        In 2004, he spent a summer at Caltech under the Summer Undergraduate Research Fellowship (SURF).

Vincent’s research is funded by the Agency of Science, Technology and Research (A*STAR), Singapore.

Vincent is a Student Member of the IEEE. Vincent is married to a wonderful woman, Huili, who is a graduate student in biology at the Whitehead Institute.

Research

My research lies in the broad areas of statistical signal processing, convex optimization and machine learning.

Current Research

I have been working on the use of convex optimization and information theory to learn probabilistic graphical models for the specific purpose of hypothesis testing/classification (SSP 2007). This work has been extended to sequentially and jointly learn increasingly complex probability models defined on graphical models for discriminating between two hypotheses (ICASSP 2008, ITA 2008). I am also interested in frame representations, sampling theory and signals with finite rate of innovation (TSP 2008).

Research in MSRC

I developed a graphical model of the immune system using Infer.NET, which allows Bayesian inference to be applied automatically to a specified graphical model. Our immune system model consists firstly of a Hidden Markov Model representing how allergen-specific skin prick tests (SPTs) and serum-specific IgE tests (SITs) change over time. By introducing a latent multinomial variable, we also cluster the children in an unsupervised manner into different sensitization classes. For 2 sensitization classes, the children who are vulnerable to allergies and have a high probability of having asthma (22%) are identified. For 5 sensitization classes, children in the first cluster, those who are vulnerable to allergies, have an even higher probability of having asthma (42%). The second part of the model involves using the inferred sensitization class as a label and 8 exposure variables in a Bayes Point Machine. Using multiple permutation tests, we conclude that the level of endotoxins and gender have a significant effect on a child’s vulnerability to allergies.

Previous Research

Previously, I was involved in developing new algorithms for privacy-preserving data mining. I examined the use of various and devised novel algorithms for the reconstruction of a distribution after a generic randomization process (MLDM 2007). I examined the accuracy and utility of using Kernel Density Resampling methods for privacy preservation (PinKDD 2007) in the context of distributed classification.

During my final year at Cambridge, I examined the effect of sparsity on underdetermined blind audio source separation (SPARS 2005). The results show that the separation performance is indeed correlated to the sparsity of the analysis coefficients of the sources in the transform domain. Our results also show that the use of overcomplete transforms does not lead to significant improvement in performance, because they fail to improve the sparsity measure.

Publications

  • NEW: Vincent Y. F. Tan, Animashree Anandkumar, Lang Tong and Alan Willsky, “A Large-Deviation Analysis for the Maximum-Likelihood Learning of Markov Tree Structures,” Submitted to IEEE Transactions on Information Theory, May 2009. [arXiv e-print]

 

  • Vincent Y. F. Tan, Animashree Anandkumar, Lang Tong and Alan Willsky, “A Large-Deviation Analysis for the Maximum-Likelihood Learning of Tree Structures,” Accepted to 2009 IEEE International Symposium on Information Theory (ISIT’09), Seoul, Korea, Jun 28 – Jul 3, 2009. [pdf] [Supp Material]

 

  • Vincent Y. F. Tan and Cédric Févotte, “Automatic Relevance Determination for Nonnegative Matrix Factorization” 2009 Workshop on Signal Processing with Adaptive Sparse Structured Representations (SPARS’09), St Malo, France, Apr 6 – Apr 9, 2009. [pdf] [Link] [Matlab_code]

 

  • Vincent Y. F. Tan and Vivek K. Goyal, “Estimating Signals with Finite Rate of Innovation from Noisy Samples: A Stochastic Algorithm,” Sampling Theory and Applications (SAMPTA’09), Marseille, France, May 18 – May 22, 2009. [pdf] (VKG Invited)

 

  • Vincent Y. F. Tan, John Winn, Angela Simpson, Adnan Custovic, “Immune System Modeling with Infer.NET” 2008 IEEE International Conference on e-Science (e-Science 2008), Indianapolis, Indiana, Dec 10 – Dec 12, 2008. [pdf] [Link] 

 

  • Vincent Y. F. Tan and Vivek K. Goyal, “Estimating Signals with Finite Rate of Innovation from Noisy Samples: A Stochastic Algorithm,” IEEE Transactions on Signal Processing, vol 56, no. 10, Part 2, Pages 5135-5146, Oct 2008, [pdf] [Link] [arXiv] [Matlab code] 

 

  • Vincent Y. F. Tan, John W. Fisher III, Alan S. Willsky, “Learning Max-Weight Discriminative Forests,” 2008 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Las Vegas, Nevada, Mar 30 – April 4, 2008, [pdf] [Link] 

 

  • John W. Fisher III, Vincent Y. F. Tan, Alan S. Willsky, “Learning Max-Weight Discriminative Forests,” 2008 Information Theory and Applications Workshop (ITA), La Jolla, California, Jan 27 – Feb 1, 2008. (JWF Invited)  [Link]

 

  • Sujay Sanghavi, Vincent Y. F. Tan and Alan S. Willsky, “Learning Graphical Models for Hypothesis Testing”, In IEEE Statistical Signal Processing (SSP) Workshop (2007), Madison, WI, Aug 26 - 29, 2007. [pdf] [Link] 

 

  • Vincent Y. F. Tan and See Kiong Ng, “Privacy-Preserving Sharing of Horizontally-Distributed Private Data for Constructing Accurate Classifiers”, Proceedings of the First SIGKDD International Workshop on Privacy, Security, and Trust in KDD (PinKDD'07), Lecture Notes in Computer Science (LNCS), Volume 4890, Pages 116-137, Springer, 2008. [pdf] [SpringerLink] 

 

  • Vincent Y. F. Tan and See Kiong Ng, “Privacy-Preserving Sharing of Horizontally-Distributed Private Data for Constructing Accurate Classifiers”, accepted by the First ACM SIGKDD International Workshop on Privacy, Security, and Trust in KDD (PinKDD 2007 held in conjunction with SIGKDD), San Jose, California, August 12-15, 2007. [pdf]  [Link] 

 

  • Vincent Y. F. Tan and See Kiong Ng, “Generic Probability Density Function Reconstruction for Randomization in Privacy-Preserving Data Mining”, In: P. Perner (Ed.): Proceedings of the 5th International Conference on Machine Learning and Data Mining (MLDM-07), (LNAI 4571), pp. 76-90, Leipzig, Germany, July 18-20, 2007. [pdf] [SpringerLink] 

 

  • Vincent Y. F. Tan and Cédric Févotte, “A Study of the Effect of Source Sparsity for Various Transforms on Blind Audio Source Separation Performance”. In Proceedings Workshop on Signal Processing with Adaptive Sparse Structured Representations (SPARS’05), Rennes, France, Nov 2005. [pdf] [sound samples] 

Reports and Thesis

  • Vincent Y. F. Tan, “Blind Audio Source Separation”. M.Eng. Final Report, Signal Processing Laboratory, Cambridge University Engineering Department, Jun. 2005. pdf
  • Vincent Y. F. Tan, “An Algorithm for Finding Equivalent Sources For a Wave Scattering Problem”. Summer Undergraduate Research Fellowship (SURF) Final Report, Applied and Computational Mathematics, Caltech, Aug. 2004. France, Nov 2005. pdf  

 Term Projects

  • Vincent Y. F. Tan, “Information Geometry Analysis of Learning Mixtures of Trees”. MIT 6.441 Information Theory, EECS, MIT, May 2008.  pdf
  • Vincent Y. F. Tan, “Learning Graphical Models using Information Criteria and Maximum Entropy Relaxation”. MIT 6.867 Machine Learning, EECS, MIT, Dec 2007.  pdf
  • Vincent Y. F. Tan, “Estimating the Parameters of a Signal with Finite Rate of Innovation from Noisy Samples: Deterministic and Stochastic Algorithms”. MIT 6.342 Wavelets, Approximation and Compression, EECS, MIT, May 2007. pdf
  • Vincent Y. F. Tan, “Stochastic Optimization of Keane's Bump Function”. CUED 5R1 Stochastic Optimization Coursework, EIST, Jun. 2005. pdf
  • Vincent Y. F. Tan, “Stochastic Processes: The Gibbs Sampler and the Straight Line”. CUED 5R1 Stochastic Processes Coursework, EIST, Jun. 2005. pdf
  • Vincent Y. F. Tan, “Newsvendors Tackle the Newsvendors Problem”. CUED 4E9: Quantitative Techniques in Operations Management, EIST, Jun. 2005. pdf

Presentations

  • “Large-Deviations for Learning Tree Structures”. LIDS Student Conference, Jan 2009. pdf
  • “Immune System Modeling using Infer.NET”. IEEE Conference on e-Science Conference Poster, Dec 2008. pdf
  • “A Graphical Model of the Immune System”. Presented at SSG Seminar, Sep 2008.
  • “A Graphical Model of the Immune System”. Microsoft Research Internship Talk, Aug 2008.
  • “Boosted Graphical Model Classifiers”. CNRS Talk, Aug 2008.
  • “Information Geometry and Mixtures of Trees”. 6.441 Information Theory, May 2008. pdf
  • “Boosted Graphical Model Classifiers”. RQE Talk, May 2008.
  • “Learning Max-Weight Discriminative Forests”. Presented at ICASSP, Apr 2008. pdf
  • “Learning Max-Weight Discriminative Forests”. Presented at LIDS Student Conference, Jan 2008. pdf
  • “Learning Graphical Models and Max-Weight Discriminative Forests for Hypothesis Testing” Presented at Lincoln Labs Seminar, Jan 8, 2008. pdf
  • “Learning Graphical Models for Hypothesis Testing”. Presented at SSG Seminar, Oct 2007. pdf
  • “Learning Graphical Models for Hypothesis Testing”. Poster presented at SSP 2007, Madison, Wisconsin, Aug 2007. pdf
  • “Privacy-Preserving Sharing of Horizontally-Distributed Private Data for Constructing Accurate Classifiers”. Presented at PinKDD 2007, San Jose, California, Aug 2007. pdf
  • “Generic Probability Density Function Reconstruction for Randomization in Privacy-Preserving Data Mining “. Presented at MLDM 2007, Leipzig, Germany, Jul 2007. pdf
  • “Estimating the Parameters of a Signal with Finite Rate of Innovation from Noisy Samples: Deterministic and Stochastic Algorithms”. 6.342 Wavelets, Approximation and Compression Term Project, May 2007. pdf
  • “Blind Audio Source Separation”. M.Eng. Project Presentation, Jun. 2005. pdf
  • “The Newsvendor Problem”. CUED 4M9: Quantitative Methods in Operations Management Final Presentation May. 2005. pdf
  • “An Algorithm For Finding Equivalent Sources For A Wave Scattering Problem”. Summer Undergraduate Research Fellowship (SURF) Final Presentation, Caltech, Aug. 2004. pdf

Classes

Relevant MIT Classes:

  • 18.465: High Dimensional Inference (Spring 2009)
  • 18.125: Measure and Integration (Spring 2009)
  • 6.980: Teaching in EECS (Fall 2008)
  • 18.101: Analysis II (Fall 2008)
  • 6.441: Information Theory (Spring 2008)
  • 18.100B: Analysis I (Spring 2008)
  • 9.520: Statistical Learning Theory (Spring 2008: Listener)
  • 6.867: Machine Learning (Fall 2007)
  • 6.241: Dynamic Systems and Control (Fall 2007)
  • 6.253: Convex Analysis (Fall 2007: Listener)
  • 6.342: Wavelets, Approximation and Compression (Spring 2007)
  • 6.252: Nonlinear Programming (Spring 2007)
  • 6.437: Inference and Information (Spring 2007: Listener)
  • 6.341: Discrete-Time Signal Processing (Spring 2004)
  • 6.432: Stochastic Processes, Estimation and Detection (Spring 2004)
  • 6.301: Solid-State Circuits (Spring 2004)
  • 6.302: Feedback Systems (Fall 2003)
  • 6.011: Introduction to Signal Processing, Communications and Control (Fall 2003)
  • 6.431: Applied Probability (Fall 2003)

 

Relevant Cambridge Classes (All 2004/2005):

Teaching

I was a Teaching Assistant for Dynamic Systems and Control (6.241) in the Fall of 2008. 

 

I gave a short review course on 6.097 Signals and Systems in IAP 2008 and IAP 2009.

 

I conducted recitations for Analytical Methods in ECE (EE2012) at the National University of Singapore (NUS) in the Fall of 2006.  

Resume

I am no longer looking for internships but here is my resume. pdf 

Contact Information

  • E-mail: vtan {at} mit {dot} edu,
  • Term Address: Room 32-D570, Massachusetts Institute of Technology
  • Telephone: (617)-913-4213

Updated: Apr 2009

 

Site Meter